Exploratory data analysis - temperature and depth

Load shark T/P data

load(file = "../data/Rdata/shark_tp_data.Rdata")
shark_tp_data

Depth overview - by individual

shark_tp_data %>% 
  ggplot() +
  geom_point(aes(x = detection_timestamp, y = Depth)) +
  scale_y_reverse() +
  facet_wrap(~ transmitter_id)

## Warning: Removed 317271 rows containing missing values (`geom_point()`).

shark_tp_data %>% 
  ggplot() +
  geom_point(aes(x = detection_timestamp, y = Depth)) +
  scale_y_reverse() +
  facet_wrap(~ transmitter_id)

## Warning: Removed 317271 rows containing missing values (`geom_point()`).

Depth overview - by receiver station

shark_tp_data %>% 
  #filter(transmitter_id == "A69-9004-14037") %>%  # If you want to look at one individual
  ggplot() +
  geom_point(aes(x = detection_timestamp, y = Depth)) +
  scale_y_reverse() +
  facet_wrap(~ station_name_long)

## Warning: Removed 317271 rows containing missing values (`geom_point()`).

shark_tp_data %>% 
  #filter(transmitter_id == "A69-9004-14037") %>%  # If you want to look at one individual
  ggplot() +
  geom_point(aes(x = detection_timestamp, y = Depth)) +
  scale_y_reverse() +
  facet_wrap(~ station_name)

## Warning: Removed 317271 rows containing missing values (`geom_point()`).

Some receiver stations failed early in the deployment period

Shark depth - daily summary

shark_daily_depths <- 
  shark_tp_data %>% 
  mutate(date = date(detection_timestamp),
         year_week = str_c(str_sub(isoyear(date), 3, 4),  "/", formatC(isoweek(date), format = "f", digits = 0, width = 2, flag = "0"))) %>% 
  filter(!is.na(Depth)) %>% 
  group_by(date) %>% 
  summarise(daily_max_depth = max(Depth),
            daily_min_depth = min(Depth),
            daily_mean_depth = mean(Depth)) %>% 
  mutate(daily_min_depth = if_else(daily_min_depth <0, 0, daily_min_depth))

#save(shark_daily_depths, file = "../data/Rdata/shark_daily_depths.Rdata")

Plot

Daily depths

depth_plot <- 
shark_daily_depths %>%   
  ggplot() +
  geom_line(aes(x = date, y = daily_min_depth), colour = "skyblue") +
  #geom_smooth(aes(x = date, y = daily_min_depth), colour = "blue",  method = "gam", formula = y ~ s(x, bs = 'cs')) + # method = "gam" implements mgcv
  geom_line(aes(x = date, y = daily_mean_depth), colour = "dodgerblue") +
  geom_smooth(aes(x = date, y = daily_mean_depth), colour = "dodgerblue",  method = "gam", formula = y ~ s(x, bs = 'cs')) +
  geom_line(aes(x = date, y = daily_max_depth), colour = "midnightblue") +
  geom_smooth(aes(x = date, y = daily_max_depth), colour = "midnightblue",  method = "gam", formula = y ~ s(x, bs = 'cs'), level = 0.95) +
  #geom_smooth(aes(x = date, y = daily_max_depth), colour = "red") + # Basic smooth, rather than gam
  theme_gray() +
  scale_y_reverse() +
  #scale_x_date(date_breaks = "1 months", date_labels = '%Y-%m') +
  #theme(axis.text.x = element_text(angle = 45)) +
  #theme(axis.text.x = element_blank()) +
  #theme(axis.title.x = element_blank()) +
  #theme(axis.ticks.x = element_blank()) +
  xlab("Date") +
  ylab("Depth \n (m below sealevel)")

depth_plot

Weekly summary

shark_weekly_depths <- 
shark_tp_data %>% 
  mutate(week_date = floor_date(detection_timestamp, "week")) %>% 
  filter(!is.na(Depth)) %>% 
  group_by(week_date) %>% 
  summarise(weekly_max_depth = max(Depth),
            weekly_min_depth = min(Depth),
            weekly_mean_depth = mean(Depth)) %>% 
  mutate(weekly_min_depth = if_else(weekly_min_depth <0, 0, weekly_min_depth))

shark_weekly_depths %>% 
  ggplot() +
  geom_line(aes(x = week_date, y = weekly_min_depth), colour = "blue") +
  geom_line(aes(x = week_date, y = weekly_mean_depth), colour = "grey30") +
  geom_smooth(aes(x = week_date, y = weekly_mean_depth), colour = "grey30", method = "gam", formula = y ~ s(x, bs = 'cs'), level = 0.95) +
  geom_line(aes(x = week_date, y = weekly_max_depth), colour = "red") +
  geom_smooth(aes(x = week_date, y = weekly_max_depth), colour = "red", method = "gam", formula = y ~ s(x, bs = 'cs'), level = 0.95) +
  scale_y_reverse() +
  xlab("Date") +
  ylab("Averaged weekly depths \n (m below sealevel)")

daymax_dep_plot <- 
shark_tp_data %>% 
  mutate(date = date(detection_timestamp)) %>% 
  filter(!is.na(Depth)) %>% 
  group_by(date) %>% 
  summarise(daily_max_depth = max(Depth),
            daily_min_depth = min(Depth),
            daily_mean_depth = mean(Depth)) %>% 
  ungroup() %>% 
  ggplot() +
  geom_point(aes(x = date, y = daily_max_depth)) +
  #geom_point(aes(x = date, y = roll_mean), data = SST_comb3, colour = "blue") +
  #geom_line(aes(x = date, y = roll_mean), data = SST_comb3, colour = "blue") 
  #geom_smooth(aes(x = date, y = max_depth)) +
  scale_y_reverse() 

Depth - by receiver location

shark_tp_data %>% 
  #filter(station_name_long == 'Lagoon_2')# %>% # Just looking at Silver City for now as it is the only one with observations for the whole period
  ggplot() +
  geom_point(aes(x = detection_timestamp, y = Depth)) +
  scale_y_reverse()  +
  facet_wrap(~station_name)

## Warning: Removed 317271 rows containing missing values (`geom_point()`).

Same with depth - pattern replicated by most individuals. Interesting to look at #14049 who didn’t make as many deep moves and whose overall temperatures look higher than the others.